CheXGround: Anatomical Region Tokens for Grounded Longitudinal Chest X-ray Interpretation
2026-08-31 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors created CheXGround, a model that helps compare chest X-rays taken at different times by focusing on specific body parts. It links these body parts in the images with the words in doctors' reports to better understand changes over time. They trained the model to match phrases in the reports with the right spots in the X-rays, improving accuracy in answering questions and describing findings. Their tests showed CheXGround does better than earlier models at understanding both the images and the language about them.
chest X-raylongitudinal modelingvisual groundingregion-of-interest (ROI)temporal reasoningvisual question answering (VQA)anatomical regionsradiology language modelpretrainingreport generation
Authors
Adonay Demewez Gebremedhin, Wessam Shehieb, Sara Alansari, Mohamad Alansari, Muzammal Naseer, Sajid Javed, Naoufel Werghi
Abstract
Recent radiology multi-modal language models have made substantial progress in chest X-ray report generation, visual question answering, and temporal reasoning. While longitudinal chest X-ray interpretation compares sequential examinations to describe change, visual grounding aims to connect clinical language with localized image evidence. Although longitudinal modeling and visual grounding have each advanced radiology language models, how localized visual evidence can support longitudinal interpretation remains under-explored. We introduce CheXGround, a region-grounded longitudinal chest X-ray language model that represents paired studies through corresponding anatomical regions. CheXGround extracts anatomical regions from current and prior radiographs, encodes them as temporally enhanced Region-of-Interest (ROI) tokens, and combines them with global temporal image context during generation. To connect these region tokens with clinical text, we propose Temporal Region--Phrase Alignment, a pretraining objective that aligns temporal anatomical representations with localized report phrases. We evaluate CheXGround on single-study and longitudinal Visual Question Answering (VQA), longitudinal findings generation, temporal grounded VQA, and anatomical grounding. Across these tasks, CheXGround improves clinical language quality, temporal reasoning, and localization accuracy over recent baselines. Our results suggest that organizing longitudinal evidence at the anatomical level is a strong representation for grounded radiology language modeling. Project page: https://adonaydem.github.io/chexground-website